Skip to main content

Music Melodic-Device

Apply chromatic and diatonic transformations to music notes

DESCRIPTION

This class provides methods to do chromatic and diatonic transformations to music notes.

The three major categories are transpositions, inversions, and five types of ornamentation.

Accidental notes are returned as "sharps" by default. To get "flats" back, set the flat attribute to True. Also, please note that the music21 library uses the hyphen (-) to indicate the flat accidental, not 'b'. Sharp is a '#', though.

SYNOPSIS

from music_melodicdevice import Device

device = Device( # defaults:
    scale_note='C'
    scale_name='chromatic'
    notes=[] # notes set and used by devices
    flat=False # return notes with accidentals as flats instead of sharps
    pattern=[0,1,2], # the arpeggiation pattern (up)
    repeats=1, # how many times to repeat the arpeggiation pattern
    verbose=False
)

# default scale: chromatic
device = Device(notes=['C4', 'E4', 'D4', 'G4', 'C5'])
notes = device.transpose(2) # ['D4', 'F#4', 'E4', 'A4', 'D5']
notes = device.invert('C4') # ['C4', 'G#3', 'A#3', 'F3', 'C3']

# diatonic transformation:
device = Device(scale_name='major', verbose=True)
device.notes = ['C4', 'E4', 'D4', 'G4', 'C5']
notes = device.transpose(2) # ['E4', 'G4', 'F4', 'B4', 'E5']
notes = device.invert('C4') # ['C4', 'A3', 'B3', 'F3', 'C3']

# unknown note:
device = Device()
device.build_scale('major')
notes = device.transpose(2, ['C4', 'E4', 'D#4', 'G4', 'C5'])
# ['E4', 'G4', None, 'B4', 'E5']
notes = device.invert('C4', ['C4', 'E4', 'D#4', 'G4', 'C5'])
# ['C4', 'A3', None, 'F3', 'C3']

# accidental notes:
device = Device(scale_note='G', scale_name='major')
device.notes = ['C4', 'E4', 'D4', 'F#4', 'C5']
notes = device.transpose(2) # ['E4', 'G4', 'F#4', 'A4', 'E5'])
device = Device(scale_note='G', scale_name='major', flat=True)
device.notes = ['C4', 'E4', 'D4', 'F#4', 'C5']
notes = device.transpose(2) # ['E4', 'G4', 'G-4', 'A4', 'E5'])

# Ornamentation:

# chromatic
device = Device()

notes = device.grace_note(1, 'D5') # [[1/16, 'D5'], [1 - 1/16, 'D5']])
notes = device.grace_note(1, 'D5', offset=1) # [[1/16, 'D#5'], [1 - 1/16, 'D5']])
notes = device.grace_note(1, 'D5', offset=-1) # [[1/16, 'C#5'], [1 - 1/16, 'D5']])

notes = device.turn(1, 'D5') # [[1/4,'D#5'], [1/4,'D5'], [1/4,'C#5'], [1/4,'D5']])
notes = device.turn(1, 'D5', offset=-1) # [[1/4,'C#5'], [1/4,'D5'], [1/4,'D#5'], [1/4,'D5']])

notes = device.trill(1, 'D5', number=2, offset=1)
# [[1/4,'D5'], [1/4,'D#5'], [1/4,'D5'], [1/4,'D#5']])
notes = device.trill(1, 'D5', number=2, offset=-1)
# [[1/4,'D5'], [1/4,'C#5'], [1/4,'D5'], [1/4,'C#5']])

notes = device.mordent(1, 'D5', offset=1) # [[1/4,'D5'], [1/4,'D#5'], [1/2,'D5']])
notes = device.mordent(1, 'D5', offset=-1) # [[1/4,'D5'], [1/4,'C#5'], [1/2,'D5']])

notes = device.slide(1, 'D5', 'F5') # [[1/4,'D5'], [1/4,'D#5'], [1/4,'E5'], [1/4,'F5']])
notes = device.slide(1, 'D5', 'B4') # [[1/4,'D5'], [1/4,'C#5'], [1/4,'C5'], [1/4,'B4']])

# diatonic
device = Device(scale_name='major')

notes = device.grace_note(1, 'D5') # [[1/16, 'D5'], [1 - 1/16, 'D5']])
notes = device.grace_note(1, 'D5', offset=1) # [[1/16, 'E5'], [1 - 1/16, 'D5']])
notes = device.grace_note(1, 'D5', offset=-1) # [[1/16, 'C5'], [1 - 1/16, 'D5']])

notes = device.turn(1, 'D5', offset=1) # [[1/4,'E5'], [1/4,'D5'], [1/4,'C5'], [1/4,'D5']])
notes = device.turn(1, 'D5', offset=-1) # [[1/4,'C5'], [1/4,'D5'], [1/4,'E5'], [1/4,'D5']])

notes = device.trill(1, 'D5', number=2, offset=1) # [[1/4,'D5'], [1/4,'E5'], [1/4,'D5'], [1/4,'E5']])
notes = device.trill(1, 'D5', number=2, offset=-1) # [[1/4,'D5'], [1/4,'C5'], [1/4,'D5'], [1/4,'C5']])

notes = device.mordent(1, 'D5', offset=1) # [[1/4,'D5'], [1/4,'E5'], [1/2,'D5']])
notes = device.mordent(1, 'D5', offset=-1) # [[1/4,'D5'], [1/4,'C5'], [1/2,'D5']])

# arpeggiation
device = Device()
notes = device.arp(['C4','E4','G4']) # [ [1/3, 'C4'], [1/3, 'E4'], [1/3, 'G4'] ]
device.pattern = [2,1,0]
notes = device.arp(['C4','E4','G4']) # [ [1/3, 'G4'], [1/3, 'E4'], [1/3, 'C4'] ]

METHODS

transpose

notes = device.transpose(amount)

Transpose a note by an integer amount.

invert

notes = device.invert(axis_note, note_list)

Invert the note_list of named notes, around the named axis_note.

grace_note

notes = device.grace_note(duration, pitch, offset=offset)

Return a list with a 64th-note grace-note and the original named pitch with octave. The grace-note is offset semi-tone steps from the given pitch. The duration is the total, original length of the pitch.

turn

notes = device.turn(duration, pitch, offset=offset)

Return a list of four notes, having an up-down pattern, in place of the given pitch and duration.

trill

notes = device.trill(duration, pitch, number=number, offset=offset)

Return a list of pairs of notes, given the number, in place of the given pitch and duration.

The first of the pair is part of the original note, and the second is the note plus the given offset.

mordent

notes = device.mordent(duration, pitch, offset=offset)

Return a list of three notes in place of the given pitch and duration.

slide

notes = device.slide(duration, from_pitch, to_pitch)

Return a list of chromatic notes inclusively between the from_pitch and to_pitch, in place of the given duration.

arp

notes = device.arp() # use defaults
notes = device.arp(notes)
notes = device.arp(notes, duration=duration)
notes = device.arp(notes, arp_type=arp_type)
notes = device.arp(notes, repeats=repeats)

Return a list of the notes selected by the arpeggiated pattern, and distributed across the duration.

MUSICAL EXAMPLES

from music21 import duration, note, stream
from music_melodicdevice import Device

device = Device(notes=['C4', 'E4', 'D4', 'G4'])
notes = device.invert('C5')

s = stream.Stream()
p = stream.Part()

for i in device.notes + notes:
    n = note.Note(i)
    n.duration = duration.Duration(1)
    p.append(n)

s.append(p)
s.show()
notes = ['C4', 'E4', 'D4', 'G4']

device = Device(scale_name='major')
device.notes = notes
device.notes = device.invert('C5')
device.notes = device.transpose(-5)

s = stream.Stream()
p = stream.Part()

for i,j in enumerate(notes + device.notes):
    if (i + 1) % 4 == 0:
        turn = device.turn(1, j)
        for t in turn:
            m = note.Note(t[1])
            m.duration = duration.Duration(t[0])
            p.append(m)
    else:
        n = note.Note(j)
        n.duration = duration.Duration(1)
        p.append(n)

s.append(p)
s.show()
from chord_progression_network import Generator
from music_melodicdevice import Device
from music21 import duration, note, stream

s = stream.Score()
p = stream.Part()

g = Generator(
    max=16,
    scale_name='whole-tone scale',
    net={
        1: [2,3,4,5,6],
        2: [1,3,4,5,6],
        3: [1,2,4,5,6],
        4: [1,2,3,5,6],
        5: [1,2,3,4,6],
        6: [1,2,3,4,5],
    },
    chord_map=['7'] * 6, # every chord is the same flavor
    substitute=True,
    verbose=False,
)
phrase = g.generate()

device = Device(verbose=False)

for i, ph in enumerate(phrase):
    arp_type = 'up' if i % 2 == 0 else 'down'
    arped = device.arp(ph, duration=1, arp_type=arp_type, repeats=1)
    for a in arped:
        n = note.Note(a[1])
        n.duration = duration.Duration(a[0])
        p.append(n)

s.append(p)
s.show()

SEE ALSO

https://en.wikipedia.org/wiki/Inversion_(music)#Melodies

https://en.wikipedia.org/wiki/Ornament_(music)

Metadata

Release files for music-melodicdevice 0.3.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for music-melodicdevice 0.3.0
File Size Uploaded Provenance
music_melodicdevice-0.3.0.tar.gz 45.0 kB Without provenance Details

Built distribution (wheel)

Table of built distributions (wheels) for music-melodicdevice 0.3.0
File Interpreter ABI Platform
music_melodicdevice-0.3.0-py3-none-any.whl Python 3 none any Details

Total release size: 75.5 kB

Release files / music_melodicdevice-0.3.0.tar.gz

Download URL music_melodicdevice-0.3.0.tar.gz
Size 45.0 kB
Tags Source
SHA-256 checksum
How to use checksums
76502f2daa3814b7f9eeacc627b826fcf3912de76d28ef3e093058f937d50d52
BLAKE2b-256 checksum
How to use checksums
341b7163a906ab1719d85b14b88cf487419ba669eae16a23d7f34fa4060908cd
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.2.0 CPython/3.13.7

Release files / music_melodicdevice-0.3.0-py3-none-any.whl

Download URL music_melodicdevice-0.3.0-py3-none-any.whl
Size 30.5 kB
Tags Python 3
SHA-256 checksum
How to use checksums
18f9d97d452fcf101360b7b9426d7352dbe27cd4a8bcaf0f665f3610685b0cdd
BLAKE2b-256 checksum
How to use checksums
3d4eb1b065236cf5dc9e938ca30e56dc58dad3428ace8a2d2e749ec16aca542c
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.2.0 CPython/3.13.7

Security for music-melodicdevice 0.3.0

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page